Comment by _heimdall

Its interesting to see what an LLM might say here, but ultimately an algorithmic prediction of how a person would answer isn't worth much.

If sources are provided and the sources check out that's one thing, but then it doesn't need to attempt to predict a likely human response to the question.

That said, as you mentioned below the note that 17% of emissions is generally attributed to commuting is relevant if true. A person staying home, requiring more energy both for lighting, HVAC, computers, etc could potentially cancel that out. Or not, and that's really my main point above as its an extremely complex situation to attempt to model and quantify.